Reinforcement Learning for Video Games

Because who needs human instincts, anyway?

Reinforcement learning, the art of training AI agents to do our work for us. In video games, it's about finding the most efficient ways to get from the couch to the fridge.

Subfields of RL4G

Policy Gradient: The most basic, the most advanced, the most overcomplicated, and most underwhelming.
Q-Learning: When the only feedback is the score.
Value Function Learning: Because who needs human emotions, anyway?
Deep RL: Because we can.
Multi-Agent RL: Because someone's gotta do the laundry, anyway.

Join the discussion, or just lurk. We won't judge. Much.